Re: [Xen-users] vCPUs and Weight/Caps

[Mohamed King <mohammed.king@xxxxxxxxx>]

Peter,

Thank you so much, that explains it.

Thank you,
Mohamed King

On Jan 13, 2013, at 1:39 PM, Peter Viskup <skupko.sk@xxxxxxxxx> wrote:

> On 01/13/2013 07:33 PM, Peter Viskup wrote:
>> On 01/13/2013 06:12 PM, Mohamed King wrote:
>>> Hi Peter,
>>>
>>> Thanks for the response. I did look at the docs prior to asking. I guess 
>>> what I am asking is what the advantages would be if you have say 2 vCPUs on 
>>> a VM.
>>>
>>>
>>> Sent from my iPad
>>
>> Hi Mohamed,
>> I understand its advantage as better possibility to schedule provided CPU 
>> time between the processes within the VM. Of course your VM will not be able 
>> to use more than 100% of one physical CPU, but should be able to run another 
>> process once one will block the first vCPU (and only in case the VM will 
>> still have enough credits on hypervisor - if credit scheduler is used). It's 
>> good to have more vCPUs on webhosting for example and on all systems where 
>> do you expect parallel data processing (webservers, databases, virtual 
>> desktop, ...).
>> This is my understanding. <correct me if I'm wrong>
>>
>> Best regards,
>> --
>> Peter Viskup
>
> Just one other thing where it is expected to have more vCPUs.
> You are not allowed to assign more cap than 100% to VM with only one vCPU. In 
> other words one vCPU can have a max. of 100% of real CPU power. That means in 
> case you need to assign more CPU time to the VM you need to setup more vCPUs 
> for that VM - based on the VM requirements (e.g. at least 3vCPUs for VM with 
> cap=280).
>
> BR,
> --
> Peter

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxx
http://lists.xen.org/xen-users